Point Clouds and Meshes
A 3D scan is not necessarily one finished model. Reconstruction may first produce a point cloud: a collection of sampled positions that describe observed surfaces. A later process can connect or reconstruct those samples into a mesh: a continuous polygonal surface.
Neither representation is universally better. A point cloud stays close to the sampled geometric evidence. A mesh is easier to display, edit and prepare for many downstream uses, but its surface also contains decisions made during meshing.
Understanding that distinction helps you inspect results realistically and choose an appropriate output.
A scan can produce several forms of geometry
A useful public model separates five stages:
- Capture evidence — video frames and other observations record the rotating object.
- Reconstruction — the selected Vision, Laser or Combo workflow estimates geometry.
- Point cloud — reconstructed surface samples are stored as individual 3D points.
- Mesh — a surface-reconstruction process connects or fits a polygonal surface to the geometric evidence.
- Prepared model — the mesh may be inspected, repaired, scaled, simplified or modified for its intended use.
These stages should not be collapsed into the claim that scanning automatically creates a fabrication-ready object. Each stage has a different responsibility, and each can preserve, remove or infer information.
A point cloud samples locations on observed surfaces
A point cloud contains many positions in three-dimensional space. Depending on the workflow and file, points may also carry color or an estimated normal indicating the local surface direction.
Point clouds are useful because they expose the reconstructed samples without requiring every gap to be filled. They support inspection, registration and Multi Merge, and they can serve as geometric reference data in other software.
They are not continuous surfaces. Zooming in reveals separation between points, and regions that were not observed may contain no samples. Point density can vary with visibility, method, filtering and reconstruction conditions.
A mesh connects geometry into a surface
A polygon mesh represents a surface using:
- vertices, which are 3D positions;
- edges, which connect vertices; and
- faces, usually triangles, which span the surface.
Meshes are widely used for viewing, editing, rendering and 3D printing. They can also carry normals, vertex colors, texture coordinates and material references.
A mesh can look more complete than its source point cloud because the meshing process creates continuity. That continuity is useful, but it must not be mistaken for directly observed detail everywhere.
Meshing interprets the evidence
Surface reconstruction must decide how nearby samples belong together. Depending on the input and settings, it may smooth noise, bridge small gaps, close a surface or create geometry across poorly supported regions.
Common issues include:
- holes where evidence is insufficient;
- false bridges between nearby but separate parts;
- softened edges or small details;
- rough regions caused by noise or outliers;
- unwanted closing surfaces near the turntable contact area;
- excessive polygon density for the intended software.
A watertight mesh is valuable for many fabrication workflows, but watertightness alone does not prove geometric correctness.
Choose the representation for the next task
| Intended task | Useful starting representation |
|---|---|
| Inspect reconstructed samples | Point cloud |
| Register and merge complementary scans | Point cloud |
| Use geometry as a visual reference | Point cloud or mesh |
| Edit a continuous surface | Mesh |
| Prepare for 3D printing | Inspected and prepared mesh |
| Preserve color appearance | A color-capable point-cloud or mesh format |
Start by asking what information the next tool requires. Then inspect whether the selected representation preserves the scale, geometry, color and surface properties that matter to the task.
Learn about output formats and post-processing
Understand preparation for 3D printingASCAND Vision, Laser and Combo scans are different ways of obtaining and interpreting geometric evidence. They are not three quality levels.